GEOREFERENCING UAS DERIVATIVES THROUGH POINT CLOUD REGISTRATION WITH ARCHIVED LIDAR DATASETS
M. S. L. Y. Magtalas
a
, J. C. L. Aves
a
, A. C. Blanco
ab a
Phil-LiDAR 2 Project 1: Agricultural Resources Extraction from LiDAR Surveys Training Center for Applied Geodesy and Photogrammetry, University of the Philippines, Diliman, Quezon City
b
Department of Geodetic Engineering, University of the Philippines, Diliman, Quezon City
KEYWORDS: UAV, Georeferencing, Registration, ICP, LiDAR ABSTRACT:
Georeferencing gathered images is a common step before performing spatial analysis and other processes on acquired datasets using unmanned aerial systems UAS. Methods of applying spatial information to aerial images or their derivatives is through onboard GPS
Global Positioning Systems geotagging, or through tying of models through GCPs Ground Control Points acquired in the field. Currently, UAS Unmanned Aerial System derivatives are limited to meter-levels of accuracy when their generation is unaided with
points of known position on the ground. The use of ground control points established using survey-grade GPS or GNSS receivers can greatly reduce model errors to centimeter levels. However, this comes with additional costs not only with instrument acquisition and
survey operations, but also in actual time spent in the field. This study uses a workflow for cloud-based post-processing of UAS data in combination with already existing LiDAR data. The georeferencing of the UAV point cloud is executed using the Iterative Closest
Point algorithm ICP. It is applied through the open-source CloudCompare software Girardeau-Montaut, 2006 on a
‘skeleton point cloud’. This skeleton point cloud consists of manually extracted features consistent on both LiDAR and UAV data. For this cloud,
roads and buildings with minimal deviations given their differing dates of acquisition are considered consistent. Transformation parameters are computed for the skeleton cloud which could then be applied to the whole UAS dataset. In addition, a separate cloud
consisting of non-vegetation features automatically derived using CANUPO classification algorithm Brodu and Lague, 2012 was used to generate a separate set of parameters. Ground survey is done to validate the transformed cloud. An RMSE value of around 16
centimeters was found when comparing validation data to the models georeferenced using the CANUPO cloud and the manual skeleton cloud. Cloud-to-cloud distance computations of CANUPO and manual skeleton clouds were obtained with values for both equal to
around 0.67 meters at 1.73 standard deviation.
1. INTRODUCTION
The significant advancement of photogrammetric processing software has transformed UAVs Unmanned Aerial Vehicles
from being recreational devices into important tools for the geospatial field. UAVs or pilotless remotely controlled aircraft
capture images from an aerial perspective. These images can further be processed to generate Digital Elevation Models
DEMs, and Orthomosaics for mapping. Without using GCPs, GPS tracking devices that can be carried by these small
instruments are limited to meter-levels of accuracy such as the popular Sensefly eBee Sensefly, 2015. Adding GCPs Ground
Control Points in processing can greatly reduce model errors to centimeter levels but these can only be acquired by doing ground
surveys in the field. For survey grade mapping of large areas, such as whole cadasters, UAS manufacturers have developed a
way of using RTK with the UAV simultaneously. As the extra equipment brings accuracies to centimeter level, the cost of
procuring the system as a whole also increases. Moreover, using any of these extra steps in the field would require more man-
power and more time. Eliminating on-site methods, this research looks at post-processing steps instead.
One alternative method is by using already existing data of relatively higher levels of accuracy and use them as reference for
the UAV derivatives. Such reference data must also have high availability to be of actual use or at least have overlaps with the
study areas to be mapped. For the case of the Philippines, a high accuracy collection of data is provided by its nationwide Aerial
LiDAR Surveys ALS. From as early as 2011, parts of the Philippines have been mapped using LiDAR Light Detection
and Ranging with derivatives being used for flood modeling and hazard mapping DREAM Program, 2012. Aside from disasters
and hazards, programs have also been conceptualized with aims for using those same point clouds in providing a detailed
inventory of the country’s natural resources Blanco et al, 2015.
In 2016, an online web portal was launched that serves as a data distribution center of these various LiDAR derivatives such as
DEMs, maps, orthophotos, and classified point clouds. These LiDAR data have accuracies of 0.5 meters on the horizontal plane
and 0.2 meters along the vertical axis. Given this collection of reference data, automated georeferencing
becomes a viable approach for UAV data. A similar concept applied to TLS or Terrestrial Laser Scanning can be used for this
study. TLS is a grounded version of ALS opting instead for a terrestrial perspective in scanning objects and environments. In
place of an aerial platform, an instrument set-up is placed to perform the scan and if need be, multiple set-ups to cover the
whole scene of interest. A simple step called Point Cloud Registration could get a whole environment in one coordinate
system by fusing datasets from multiple views Rajendra et al, 2014. Cloud registration is performed along overlapping areas
of different scans and as those areas of overlap become consistent with each other, so do the rest of the dataset. The parameters used
for this process include translation and rotation. This ensures that the cloud transforms only with location and distance.
2. METHODOLOGY
